For some reason known only to the collective marketing mind of the ICT industry, business intelligence (BI) has managed to become more than a little detached from the analytics that underpin it. In business generally, BI today is seen principally in its manifestation as dashboards, traffic lights and other pretty visuals to indicate how key performance indicators (KPI) currently stand. In other words, it is well advanced along the hype cycle. But the dashboard metaphor continues to be a good one. We can make driving decisions based on speed, RPM, fuel status and so on without having much of a clue about what goes on under the bonnet. Is the speedometer mechanical or electronic? Nobody needs to know. So long as we can have confidence that the dashboard dials and displays accurately show what is happening we can use and depend on them to drive well and to drive safely.
Very much the same is true of BI. Under the bonnet there is a great deal going on and what you see on your dashboard is all you need for practical purposes. So users generally think whatever tools they have bought are BI without realising that it is in fact not a technical term at all but a generally agreed description of the software tools that organisations can use to search their combined databases and find and present usable information. Its origins, amongst others, are in data mining where patterns can be discovered in multiple types of data and data stores.
The principal purpose of BI in business is to provide analyses of underlying data that are easy to understand and act on. That is why the user front end of every BI system will be some form of visual representations of the analysis that has been performed. But one danger already becoming apparent in the market is that the deliberate simplicity of the product, the attractive and accessible graphic display, obscures the depth and technical complexity of the data analysis that must underpin it. That in turn may be leading many organisations looking at BI to query the investment and the effort required. The product is simple-the process is far from it.
The ability to use quantitative analytical data to shape decisions and outcomes has become a key source of competitive advantage over the past decade or so, according to Brian McCarthy, who has a global role as strategy director in Accenture Analytics, and is one of the authors of its recent report “Getting Serious about Analytics to Drive Outcomes.” Accenture has been involved with business analytics for over three decades, he points out, and late last year announced the setting up of the Accenture Analytics Innovation Centre in Dublin with a recruitment target of 100 specialists. Accenture research confirms that high performance businesses, those that substantially outperform competitors over the long term and across economic, industry, and leadership cycles, are five times more likely to use analytics strategically compared with low performers.
What is happening today at the top level, McCarthy says, is that business analytics technology has moved up the sophistication curve and has advanced beyond delivering descriptive and current information towards increasingly powerful predictive analytics. “It is all about being able to answer questions like ‘What would happen if…?’ and modelling a range of scenarios. All enterprises have to deal with business and market volatility of varying kinds. Predictive analytics takes us from the operational level, which is where BI and dashboards and so on are currently monitoring performance, up through the tactical to the strategic level.”
McCarthy goes further to say that the next big wave in business analytics will be ‘decision process re-engineering.’ “We will also see migration from the narrow functional focus to cross-functional and horizontal. Management has to have that view right across the enterprise because, in essence, the challenge is to measure what matters-when what matters is constantly changing!” There is an accelerating trend towards automated decision making, monitoring key metrics but then sensing specific information that will cause something different to happen or trigger human intervention. At the same time the capability is in place to measure the effectiveness of the action and to feed back into the decision making processes.
Self service
Microsoft has a vision of self-service BI, based on familiar tools, which will give the user multiple ways of analysing data. Looked at this way, much of the value of the BI is that it is inherently collaborative, based on SharePoint and PerformancePoint. Integrated analytics help people to monitor KPIs and other information relative to their own responsibilities. They can go on to analyse it further and easily share the results and insights with colleagues. “Corporate performance management is based on real data and the collective efforts across functions and teams,” says Barry McMahon, solutions and applications platform manager, Microsoft Ireland. “That tells everyone exactly what is going on relative to plans and forecasts. The next thing is to see if you can see and understand trends. It is as old as business itself-you are trying to see what is working well so you can do more of it and what is not performing so that you can change it.”
What is different today, McMahon says, is that individuals can be ‘data explorers’ and perform sophisticated analysis using desktop tools as familiar as Excel. Personal or self-service BI with SQL Server PowerPivot enables flexible access to multiple sources of data with no or minimal reliance on the IT department. “Users can also mash up data from internal and external sources, analysing vast amounts of data directly in Excel in a very intuitive and visual way. Then they can easily share their applications with others through SharePoint.”
He stresses that the concept is of managed self-service, so that information access is governed by Active Directory and identity control, for example. Individuals can use analysis tools and gain valuable insights and ultimately compress decision cycles. On the other hand, essential corporate tools and management information will be managed by IT for quality and consistency but shared in the same ways across the organisation.
“There is a wide spectrum between the kinds of self-service BI we are talking about and high level corporate BI in the management information space where the integrity and lineage of the data may have financial or regulatory implications,” McMahon says. “That is at a different pole and of course it requires IT department expertise for standards and governance. But the daily self-service BI tools, now advanced enough to allow scenario modelling and ‘What if…?’ questions, have become enormously valuable, largely because they are user-friendly and practical.”
IBM has been a leader in analytics for decades, with a long pedigree in data mining and more recently its acquisition of and investment in Cognos. “The key point about Cognos is that it can access virtually any corporate data source, regardless of the platform or applications involved, and provide all users with understandable and detailed information,” explains Liz McFadden, leader of the IBM Global Telco Business Analytics and Optimisation Centre of Excellence in Dublin. “It combines access to all of those types of data-from formal data warehouses through ERP systems and daily operational data to external data sources. Users can also shift readily from relatively simple viewing and analysis to advanced ‘What if…?’ and predictive analysis through the same interface.”
A great deal of the value to non-specialist users of BI today is the ability to see and absorb information in an easy and highly visual way, she says, with dashboards and scorecards now common across all types of organisation. “Under the covers, however, there are a lot of very smart analytics going on. Users can drill down, for example, to operational trends and thresholds and see trending patterns as they happen.” She cites customer behaviour as one area where the potential value of BI is very clear: “In business we all talk about the single, 360 degree view of the customer as the ideal. But it is an ideal and in practice we have to live with data that comes from different applications and sources, so knowing what the trusted view is can be critical. That is where the management of master data, ensuring the consistency of the core and most trusted information, is an essential layer in any serious corporate BI and analytics.”
Modelling is where analytics can become uniquely valuable to the enterprise today, Liz McFadden says. “You can combine historical data and multiple sources to show relationships and bring out the insights-even adding in real-time data streams. This is how modelling becomes predictive and potentially richer by adding more sources and parameters while revising your models in the light of events and experience.”
John Coleman, managing director of ProStrategy Colman, believes that business analytics and modelling today has become something like a multi-dimensional spreadsheet. “It can be as easy to learn and use as a spreadsheet but with a huge range of potential inputs.” Working with both IBM Cognos and Microsoft, he and his team have been involved in many BI projects over the years and have seen the technology improving and the appetite of clients for smart systems increase as the pressures on business mount.
Acknowledging the trend towards various kinds of user self-service in BI and the value of distributing analytical tools, Coleman said there are still some dangers. “The important thing is that BI is built on a platform where the data is validated and then modelled properly. Once an enterprise is large enough to invest in serious BI, say from medium sized companies upwards, I believe that data modelling has to be professional.”
After that, on the other hand, users can have a very powerful yet easy to use set of tools to inform their decisions. “It is by no means automatically a huge investment, either. I would say that something of the order of EUR*20,000 or a bit more would see a medium sized company with perhaps 50 to maybe 200 staff very well positioned to monitor its performance and engage in what is really very advanced scenario planning with Cognos Express.” That capability has been greatly enhanced recently by in-memory technology that allows the processing of very large data sets very quickly, he points out. “The actual computing power needed is not excessive, say a dual core processor running a 64-bit OS with maybe 4 to 8gigs of RAM. People are using laptops with that sort of performance-and gamers might think it quite modest!”
Fusion power
A quite different type of BI solution is the cloud-based offering from Irish company ASPeon.com which describes itself as ‘SaaS powered business intelligence.’ Its Fusion software suite is modular, with the modules combining with each other and with a client’s existing applications, according to Dave Feenan, business development director, ASPeon. “In essence, our software enables the business to give its managers up to the minute business intelligence, delivered in the first instance as dashboard presentations of whatever KPIs it is choosing to monitor.” As with almost all cloud and SaaS services, the charges are per month per user per module and are accessible from any web-enabled device. “It’s fair to say that the economic situation has accentuated the need for smart analytical tools in business and the costs are simply not an obstacle any more. There are high end BI systems that are appropriate for large organisations and their budgets and require professional installation. Our system offers a lighter weight solution for the same business needs-the ability to quickly identify emerging trends and take corrective action before the business is adversely impacted or to leverage opportunities before your competitors.”
In essence, ASPeon’s Fusion products are delivered as a service and enable organisations to use their own information better without infrastructural and implementation issues. “DashBOARD gives senior executives and departmental managers a visually appealing and easy to understand insight into important metrics and KPIs, focussed on their own spheres of responsibility,” said Feenan.
The next generation of BI and business analytics may already be here in the form of Greenplum, a specialist in large scale data warehousing and analytics. Tellingly, it was bought by EMC earlier this year and now forms the basis of EMC’s new Data Computing Division. Its star product is the Greenplum Data Computing Appliance, a highly scalable, parallel data warehousing appliance that integrates database, compute, storage and network into a single enterprise class system. Designed for rapid analysis of data volumes scaling into petabytes, the Greenplum ‘appliance’ is a powerful platform for unifying business intelligence and advanced analytics across large enterprises. The company’s web site calls it ‘The Answer Machine – Data in, Decisions out’ which is certainly a statement of marketing intent and a management dream.
Despite what the ‘appliance’ product might suggest, EMC and Greenplum share a vision of cloud computing as the major part of the answer to the requirements of analysis in dealing with ‘Big Data’-terabyte or exabyte data sets that are very difficult to work with using current software tools. Robert Klopp, EMEA technical architect in Greenplum, explains that massively parallel computing is the only technically feasible approach to modelling those enormous data sets. “The software has to deal with the entire data set for the algorithms to model it and extract the answers. Other current techniques typically model samples of the data and use data in-memory with ultra-fast processors to achieve the speed to be practical in business operation.”
The Greenplum approach, said Klopp, is to use essentially simpler mathematical algorithms to perform the analysis but eliminating ambiguity, and so producing more ‘accurate’ or realistic models, by modelling all of the data all of the time. “We are also aiming at making such analysis easier for real life business use-users will not need a PhD in maths!”
Subscribers 0
Fans 0
Followers 0
Followers